does anyone use HEATPIPES?

[roid]

Years ago i remember reading about heatpipes and how they were neat. Their use on laptops is expected nowdays.

It must be better than watercooling - completely sealed and no moving parts. So does anyone use heatpipe technology in their desktops? i'd love to see pics

[FunkyStickman]

Almost all high-end CPU coolers use heatpipes nowdays. Zalman makes several nice passive graphics coolers with them (I have one) and they even make totally silent cases that use heatpipes for CPU and graphics that connect with the side of the case.

[Krom]

It is a myth that they are more effective or 'better' than watercooling. They are still limited by the surface area of the heatsinks they are attached to and the airflow over the heatsink, just like most water cooling systems they still depend on fans to get the work done. Watercooling is so effective because you can easily attach a extremely large heatsink (radiator) to them and they transport the heat directly outside the case.

Unlike watercooling using fanless cooling towers, going with a fanless design using only heatpipes is impractical since it would require a equally massive heatsink (passive cooling towers are huge) and would also have to be very near the heat source. It is hard for a heatpipe design to vent the heat outside the case like a water cooler can.

[Vindicator]

Here's a pic of what my heatsinks look like when mounted to my dual Xeon board. You can see the pair of heatpipes running the length of the tower. For a size reference, one 80mm fan fits on the side pretty nicely. They work well but aren't capable of cooling my Xeons passively.

[Duper]

holy........

the new world trade centers.. you leasing space in those??

[Immortal Lobster]

The stock HSF with my opteron utilizes 6 heat pipes, but its still short and squaty, very effective, but I still prefer my Blue Orb II, keeps my OCed opty a stifling 6°C over ambiant, or 33.4°C, not bad for a 700mhz OC

http://img.photobucket.com/albums/v105/ ... AG0231.jpg
http://img.photobucket.com/albums/v105/ ... AG0301.jpg
Kromedit: your images are too large, try to keep them at or under 800x600 so they don't stretch the page please.

[roid]

It is a myth that they are more effective or 'better' than watercooling. They are still limited by the surface area of the heatsinks they are attached to and the airflow over the heatsink, just like most water cooling systems they still depend on fans to get the work done. Watercooling is so effective because you can easily attach a extremely large heatsink (radiator) to them and they transport the heat directly outside the case.

but how/why is this different to a heatpipe? How would watercooling be able to utilise a larger heatsink than heatpipes? i don't see why this would be so.

Unlike watercooling using fanless cooling towers, going with a fanless design using only heatpipes is impractical since it would require a equally massive heatsink (passive cooling towers are huge) and would also have to be very near the heat source.

why

It is hard for a heatpipe design to vent the heat outside the case like a water cooler can.

why

if this is all on some website that'd be cool.

Here's a pic of what my heatsinks look like when mounted to my dual Xeon board. You can see the pair of heatpipes running the length of the tower. For a size reference, one 80mm fan fits on the side pretty nicely. They work well but aren't capable of cooling my Xeons passively.

are those the 4 copper pipes sticking out of each tower? The cooling tower might work better if you put it on it's side, for convection.
Where did you get the heatpipes from, where they part of the whole cooling tower package thingy? does it have a brandname so i can look it up online?

The stock HSF with my opteron utilizes 6 heat pipes, but its still short and squaty, very effective, but I still prefer my Blue Orb II, keeps my OCed opty a stifling 6°C over ambiant, or 33.4°C, not bad for a 700mhz OC

are these heatpipes in any of your pics? i'd liek to see them

[Genghis]

Here's a review of the one I use:

http://www.silentpcreview.com/article251-page1.html

[Krom]

It is a myth that they are more effective or 'better' than watercooling. They are still limited by the surface area of the heatsinks they are attached to and the airflow over the heatsink, just like most water cooling systems they still depend on fans to get the work done. Watercooling is so effective because you can easily attach a extremely large heatsink (radiator) to them and they transport the heat directly outside the case.

but how/why is this different to a heatpipe? How would watercooling be able to utilise a larger heatsink than heatpipes? i don't see why this would be so.

Because in a watercooler the water moves through the tubes driven by a pump and a water cooling system is a loop, heat pipes do not circulate. Other then being called "pipes" water coolers and heatpipes have next to nothing in common. Heatpipes are closed non-looping systems; water coolers are looping systems. By nature the distance a heatpipe can efficiently move the heat from the original heat source is limited.

Unlike watercooling using fanless cooling towers, going with a fanless design using only heatpipes is impractical since it would require a equally massive heatsink (passive cooling towers are huge) and would also have to be very near the heat source.

why

Same reason as up there, heat pipes don't circulate like a water cooler, so it has to stay very close to the heat source. Heat pipes can only transfer heat so far so the heat pipe design can't physically reach outside the case efficiently, and there would be no room inside the case for such a massive heatsink.

It is hard for a heatpipe design to vent the heat outside the case like a water cooler can.

why

Again same reason, it can't reach like a water cooler can, my radiator is actually in a different room entirely from my PC, it is impossible to do that with a heatpipe design.

[Vindicator]

Here's a pic of what my heatsinks look like when mounted to my dual Xeon board. You can see the pair of heatpipes running the length of the tower. For a size reference, one 80mm fan fits on the side pretty nicely. They work well but aren't capable of cooling my Xeons passively.

are those the 4 copper pipes sticking out of each tower? The cooling tower might work better if you put it on it's side, for convection.
Where did you get the heatpipes from, where they part of the whole cooling tower package thingy? does it have a brandname so i can look it up online?

The case is normally mounted upright. Here's a review of those heatsinks. Note that they look a little different than in the picture; mine look like the ones in the review (the pic I posted wasnt mine but has the same hardware).

[Immortal Lobster]

are these heatpipes in any of your pics? i'd liek to see them

nope, but I can take some of them

http://img.photobucket.com/albums/v105/ ... AG0339.jpg
http://img.photobucket.com/albums/v105/ ... AG0340.jpg
Kromedit: your images are too large, try to keep them at or under 800x600 so they don't stretch the page please.

[roid]

Doesn't anyone use heatpipes to pipe heat to outside of their systems into external radiators/heatsinks? You know... just basically replacing the whole watercooling setup with a heatpipe - but retaining the large external heatsink/heatsinks.

hmm, maybe heatpipe tech hasn't yet got the mainstream acceptance of watercooling tech.

It is a myth that they are more effective or 'better' than watercooling. They are still limited by the surface area of the heatsinks they are attached to and the airflow over the heatsink, just like most water cooling systems they still depend on fans to get the work done. Watercooling is so effective because you can easily attach a extremely large heatsink (radiator) to them and they transport the heat directly outside the case.

but how/why is this different to a heatpipe? How would watercooling be able to utilise a larger heatsink than heatpipes? i don't see why this would be so.

Because in a watercooler the water moves through the tubes driven by a pump and a water cooling system is a loop, heat pipes do not circulate. Other then being called "pipes" water coolers and heatpipes have next to nothing in common. Heatpipes are closed non-looping systems; water coolers are looping systems. By nature the distance a heatpipe can efficiently move the heat from the original heat source is limited.

Unlike watercooling using fanless cooling towers, going with a fanless design using only heatpipes is impractical since it would require a equally massive heatsink (passive cooling towers are huge) and would also have to be very near the heat source.

why

Same reason as up there, heat pipes don't circulate like a water cooler, so it has to stay very close to the heat source. Heat pipes can only transfer heat so far so the heat pipe design can't physically reach outside the case efficiently, and there would be no room inside the case for such a massive heatsink.

Heatpipes arn't solid cores of copper. They are hollow pipes filled with a working fluid and a wicking system, they use phase change as their mechanism. They have equivalent thermal conductivity of over 80x solid copper (a few sources have indicated their conductivity as thousands of times that of solid copper, i'm still researching...).

They DO circulate, liquid at the vapouriser end turns to vapour and travels to the condensor end where it condenses, then it is returned to the vapouriser via capilary action of the internal wick. That is circulation - vapour goes one way, turns into liquid. liquid goes the other way, turns into vapour. repeat.

It is hard for a heatpipe design to vent the heat outside the case like a water cooler can.

why

Again same reason, it can't reach like a water cooler can, my radiator is actually in a different room entirely from my PC, it is impossible to do that with a heatpipe design.

I just spent the day reading about heatpipes.
They reach as far as you want them to. The vapour inside travels until it condenses on the cold end. The limiting factor in a heatpipe's design is the internal wick's ability to move the condensed liquid back to the hot end to be revapourised. There are many wick designs. FiberSpiral wicks don't perform very well unassisted by gravity, and Groove wicks are even worse. But both Mesh and Powder Metal wicks are relatively uneffected by the length of the heatpipe.

This may be blunt, but you don't seem to know anything about heatpipes Krom. They work on similar principles to fridges, phase change.

I'm interested in using them as an alternative to watercooling, to pipe heat to a large external (possibly passive) heatsink. Basically just replacing the waterblock, pump, and waterhose/connections with a heatpipe.

So far everything i've read indicates their very real advantages over watercooling: No moving parts (no pump) so nothing to break down. Unlimited length unconstraigned by pump power. Completely sealed, corrosion and algae growth is not an issue. The cooling mechanism is powered by the heat itself, no extra power (or noise) is needed for a pump.

I'm still trying to track down real world C/W comparisons with air and watercooling. That's what i was hoping to find here.

*sigh* i was hoping other people would know more about heatpipes.

[fliptw]

why use expesnive metal when cheap plastic tubing is much cheaper and lighter?

Besides, you can't engineer a more cost-effective solution than large volumes of water.

heatpipes need to be sealed for their life time, and with certian material combinations the liquid will diffuse thru the pipe walls over time.

[Immortal Lobster]

I knew how they operated =)

[Krom]

Actually I did read how heat pipes operate, and if making them reach further than inside a PC case was practical someone would have done it by now. The evaporative cooling method that heatpipes use is simply more efficient then using solid metal to conduct heat away, but otherwise it is identical in behavior and limitations to a solid metal heatsink. The greater efficiency allows a heatpipe to reach a little further then solid metal, which is why you see heatpipe heatsinks are larger and cool better then solid metal, but it can't beat flowing water for overcoming distance.

Also, if heatpipes get too hot they will stop working, because if all of the coolant inside a heat pipe is evaporated the C/W will be the same as a hollow copper tube.

Just for fun I looked up my CPU's wattage at stock speed (50 watts, http://www.amd.com/us-en/assets/content ... /25175.pdf ) and then calculated the overclocked wattage of my CPU using: Overclocked wattage = stock wattage * ((overclocked frequency / stock frequency) * ((overclocked voltage / stock voltage)^2))
And got 118 watts. (Whoa, and people say prescott is hot!)

According to my motherboard sensor and room thermometer; CPU: 37C, ambient 24C.
C/W = [max CPU temperature - ambient temperature] / processor wattage.
And got a C/W ratio of 0.11

[Immortal Lobster]

MY CPU sucks down about 110Watts(stock) from my previous calculations, my OC brings it to 138W, not bad for a DC

[Genghis]

C/W comparisons don't tell the whole story. You can look at the various watercooled and aircooled (which use heatpipes) charts here and it looks like the heatpipes have water cooling beat. But in reality, they just can't move as much heat as water cooling. Overclockers always get better results with water than heatpipes.

http://overclockers.com/articles373/

Heat pipes are still pretty darn good, though. Unless you're super extreme or nerdy, you don't need water for a decent overclock. I'm oc'ing my pre-conroe Intel 930D by 33% with the Ninja Scythe.

Furthermore, with the low wattage of Conroe there's one less reason to buy water kits. Of course, hardcore overclockers will always use water, but us casual overclockers don't need it anymore.

[roid]

Krom you sound like a longtime watercooling fanboy (hardcore 1337 overclocker). Which would be handy if i wanted to compare watercooling systems against eachother, but i'm just trying to rule out passive cooling (without pumps) before i dive back into watercooling.

It's true i'm reading a lot of opinions around the place that watercooling cools better. But i havn't seen anything other than opinions, i'm still searching for someone who's actually tested this ★■◆● and got some comparative figures. There's a lot of nerds out there and i'd trust their fanboyish opinions about as fast as i'd enter into a fanboy game-console discussion. If heatpipes are worse than watercooling then i'd like to know by how much because maybe i can just use bigger heatpipes! It's a design issue: silence, reliability. Heatpipes seem more reliable than watercooling - they arn't as prone to breaking down. (i dunno, last time i was reading about watercooling, they needed regular attention/maintenance - has this changed?).

IMO watercooling is overkill, even heatpipes are overkill. But compared to the disadvantages of watercooling, heatpipes are a no brainer - a "put it in and forget" solution - especially if i can pull it off with a passive radiator/heatsink. Pumps are a worry for me.

if making them reach further than inside a PC case was practical someone would have done it by now.

From what i've been reading they are well suited for such a task, what i'm trying to do is see how mainstream this is. I havn't been able to turn up much, so i'm assuming it's not yet mainstream. (but don't you remember back when watercooling was purely the domain of insane overenthusiasts?)

if heatpipes get too hot they will stop working

yep, if the heat input is beyond the boiling temperature of the working fluid. But in that situation you've probabaly got bigger problems on your mind as (assuming WATER is the working fluid) your CPU is already over 100C. So the heatpipe stops working after your CPU is already fried, not really much of an issue NE?

The greater efficiency allows a heatpipe to reach a little further then solid metal, which is why you see heatpipe heatsinks are larger and cool better then solid metal, but it can't beat flowing water for overcoming distance.

Heh, a LITTLE further? You don't think you're being a tad conservative there? 80x further, minimum!

From what i've read it seems that heatpipes used in integrated HSFs (like the ones posted here so far) are cheap and nasty - While they do still outperform solid copper, they mostly just use them for the 'woa cool' factor to move them off shelves. Such cheap and nasty integrated 'cool factor' heatpipes arn't competing with watercooling. As i said, i'm still trying to find a good comparison.

Guys, although it is interesting to see how heatpipes are currently being commonly used, and i DID ask to see where you guys were using heatpipes... But really not looking for HSFs with 'wow-factor' heatpipes integrated. I'm talking about QUALITY heatpipes to move heat outof the case to a large external radiator/hsf just like watercooling but without the pumps, leaks and algae.
I guess this is more of a CASE design. My interest was pricked when a few days ago i stumbled across an ultra-small case that used heatpipes to move the CPU HSF to where they was room - and i got thinking: Why not just use it like they use watercooling and pipe it all the way outside of the case to a huge passive radiator? Then i did some reading on heatpipe technology, and decided that it was indeed possible. So then i'm where i am now... trying to find such setups.

[Vindicator]

This may interest you, roid. http://www.silentpcreview.com/article302-page1.html

[Grendel]

IMO watercooling is overkill, even heatpipes are overkill. But compared to the disadvantages of watercooling, heatpipes are a no brainer - a "put it in and forget" solution - especially if i can pull it off with a passive radiator/heatsink. Pumps are a worry for me.

Pumps are not worse in MTBF than a high quallity fan. Even better if you get a HQ pump. A heatpipe isn't the solution to everything, it's just a means to get the heat to another location allowing you to get rid of it in a more efficient manner. You still need active cooling w/ todays high end PCs no matter what. WC isn't overkill at all, it allows you to have a quiter PC, cooler PC or OC the crap out of it, as you like. Plus you don't have to run 100's of CFM through the case, making the setup maintainance friendlier even.

if heatpipes get too hot they will stop working

yep, if the heat input is beyond the boiling temperature of the working fluid. But in that situation you've probabaly got bigger problems on your mind as (assuming WATER is the working fluid) your CPU is already over 100C. So the heatpipe stops working after your CPU is already fried, not really much of an issue NE?

That's a misconception -- the fluid in the pipe transports the heat by being "boiled" on the hot end (phase change one). The vapor then travels to the cold end and condensates (phase change two). The wick sucks up the fluid and transports it back to hot end. The fluid inside the pipe has a lot lower boiling point (like 30-40C) than water. If the whole thing heats up to that boiling point (eg. the fan on the cold end fails) the heat transport breaks up.

[Tricord]

It's all thermodynamics.. Your CPU generates heat, it has to be cooled down. If you use a passive heatsink or a heatpipe that stays inside the case, the heat will never leave the case. Some power amps use passive heatsinks that are built into the case exterior, transporting the heat directly outside the case. A passive system needs a heat outlet outside the case, because there is no transportation for heat inside the case.

Traditional fan cooling uses air as transportation factor to get the heat out. Water cooling uses water (which can transport much more heat energy per unit of volume than air). The advantages of watercooling are twofold:
1) it can take away much more heat energy from a given surface than air can and
2) it is easy to transport over relatively small distances.

I'm not sure about heatpipes, but from what I read in this thread their efficiency depends highly on their orientation and shape. If the condenser is located lower than the heat source (i.e. the pipe runs or bends downward) I can't see how a passive heat transport can be established using the evaporate/condense principles. At any rate, they can never be as effective as a closed looping system with separate flowpaths for hot and cold heat transportation media.

[roid]

THANK YOU VINDI YOU ARE AWESOME.

I'm not sure about heatpipes, but from what I read in this thread their efficiency depends highly on their orientation and shape. If the condenser is located lower than the heat source (i.e. the pipe runs or bends downward) I can't see how a passive heat transport can be established using the evaporate/condense principles. At any rate, they can never be as effective as a closed looping system with separate flowpaths for hot and cold heat transportation media.

they use capillary action, various styles of internal wicks.

[Krom]

Well I have been using water cooling almost exclusively on my main PC since sometime in 1997. My first water block was actually cut out of a 4\" PVC pipe cap and had the water flowing directly on the top of the heat spreader on an AMD K6.

On reliability: I have only ever used two pumps, the 80 GPH one I am using now is about 5 years old and has been running more or less 24/7 the whole time, its MTBF is 100,000 hours (11.4 years). The original pump I used was 60 GPH and is sitting unused but perfectly functional in the basement; I bought the pump I use now because it has threaded inputs and outputs, but both are aquarium/fountain pumps. The pumps have outlived about 2 Northbridge fans, 3 CPU fans, 2 video card fans, a couple case fans and 4 whole computers. And you can't hear either pump running over the hard drive unless you press your ear on the side of the reservoir and listen directly.

I bought my video card used from Matrix, it originally came with a heatpipe design on it, it transported the heat from a copper plate attached to the GPU on the bottom side of the card about 4 inches through a 180 curve to a heatsink on the top side of the card. When Matrix replaced it with an all copper zalman GPU heatsink fan he recorded a consistent 20-30C drop in temperature over the stock heatsink, along with improvements in overclocking.

Your idea of using heatpipes to form a large passive cooling tower won't work as well as water cooling because heatpipes can't spread the heat out over that wide of an area as effectively. Look at the case that Vindicator posted, at load the CPU (A64 3000+, only 44.1 watts TDP) reached 51C, acceptable for pure passive cooling, but it is easily matched or even outperformed by the elementary all-aluminum AMD retail heatsink fan. And my CPU at a little less then three times the wattage runs much cooler, though I can guarantee there won't be any out of the box PC water coolers that can match the performance of mine. If you want to use a large volume cooling tower, the only thing currently that can effectively spread and dissipate the heat over such a wide area is a water cooler.

I should mention that on my server computer I am able to cool its CPU at full load overclocked to within 2-3C of where it ran with the water cooler by using a Alpha PAL8045 heatsink with an 80 MM, 80 CFM, 57 dBA, 6000 RPM Delta fan. It is possible for heatpipes and heatsinks to match the performance of water cooling, but not at fewer than 50 decibels.

[Immortal Lobster]

Id like to investigate water cooling down the road, but the few obstacles I see, money, and the fact that if a pump goes out, those little waterblocks are going to be worthless as ★■◆● when it comes to temporarly passivly ooling the system whilst I scramble for the off button. hose ruptures are pretty too im told.

Water cooling is only really beneficial when overclocking and forceing way more then stock voltage to a CPU, why, becuase of the focused hot spot, at that point there is no way a passive cooler, or a heatpiped cooler, either with or without fan, are going to keep that overvolted CPU under control. Im only pushing about 5% through my vcore, up to 1.425 from 1.35, I wouldnt got much further then 1.5 on my current cooler, maye 1.525. with water, 1.75, hell 1.8 probly wouldnt scare me.

[Krom]

Like I said, the MTBF on your average water cooler pump is 11 years, any decent pump you get will likely outlive your computer a couple times over. My water block has a rather large lump of solid copper in it (over 500 grams IIRC, a \"heavy\" heatsink is 300 or more grams), if you shut off the pump a couple minutes later the CPU overheat protection kicks in and shuts down the system as the CPU temp creeps up. Even without the pump, it still has to warm up that lump of copper, and the water that remains in the water block. You have plenty of time to shut down your computer normally if you see the pump stop.

My cousin used a water cooler also, but for whatever reason he shut down the pump when he wasn't running the computer, so once his mom turned the computer on and ran it for several minutes without the pump running before the computer crashed. It boiled the water, melted the hoses, and even melted some of the solder on the water block. Once my cousin replaced the hoses, checked the block for a seal and turned the pump back on, the computer booted up normally and has been running just as overclocked ever since with no problems.

I have also never had a hose rupture in the 9 years I have been doing this, and I have no idea what idiot would run a pump that could put out enough pressure to rupture a hose in the first place.

Pump failure is not all the doom and gloom that air cooled people think it is.

[Zantor]

It is a myth that they are more effective or 'better' than watercooling. They are still limited by the surface area of the heatsinks they are attached to and the airflow over the heatsink, just like most water cooling systems they still depend on fans to get the work done. Watercooling is so effective because you can easily attach a extremely large heatsink (radiator) to them and they transport the heat directly outside the case.

but how/why is this different to a heatpipe? How would watercooling be able to utilise a larger heatsink than heatpipes? i don't see why this would be so.

Unlike watercooling using fanless cooling towers, going with a fanless design using only heatpipes is impractical since it would require a equally massive heatsink (passive cooling towers are huge) and would also have to be very near the heat source.

why

It is hard for a heatpipe design to vent the heat outside the case like a water cooler can.

why

if this is all on some website that'd be cool.

Here's a pic of what my heatsinks look like when mounted to my dual Xeon board. You can see the pair of heatpipes running the length of the tower. For a size reference, one 80mm fan fits on the side pretty nicely. They work well but aren't capable of cooling my Xeons passively.

are those the 4 copper pipes sticking out of each tower? The cooling tower might work better if you put it on it's side, for convection.
Where did you get the heatpipes from, where they part of the whole cooling tower package thingy? does it have a brandname so i can look it up online?

The stock HSF with my opteron utilizes 6 heat pipes, but its still short and squaty, very effective, but I still prefer my Blue Orb II, keeps my OCed opty a stifling 6°C over ambiant, or 33.4°C, not bad for a 700mhz OC

are these heatpipes in any of your pics? i'd liek to see them

The AMD forums have a section called AMDEdge which allowed users to learn about new technologies and techniques. The AMD forums are located at http://forums.amd.com.

[Immortal Lobster]

and since AMD Edge is more or less discontinued, lets just link them to the orignal article

linky